experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Fix ValueError("None values not supported.") in the Abalone tutorial.

Browse Source 
PiperOrigin-RevId: 161143481
This commit is contained in:
Igor Saprykin 2017-07-06 15:59:14 -07:00 committed by TensorFlower Gardener
parent f9c9cacb06
commit eb8bebd670
2 changed files with 19 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
tensorflow/docs_src/extend/estimators.md
View File

@ -578,7 +578,12 @@ def model_fn(features, labels, mode, params):
# Reshape output layer to 1-dim Tensor to return predictions # Reshape output layer to 1-dim Tensor to return predictions
predictions = tf.reshape(output_layer, [-1]) predictions = tf.reshape(output_layer, [-1])
predictions_dict = {"ages": predictions}
# Provide an estimator spec for `ModeKeys.PREDICT`.
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(
mode=mode,
predictions={"ages": predictions})
# Calculate loss using mean squared error # Calculate loss using mean squared error
loss = tf.losses.mean_squared_error(labels, predictions) loss = tf.losses.mean_squared_error(labels, predictions)
@ -594,9 +599,9 @@ def model_fn(features, labels, mode, params):
train_op = optimizer.minimize( train_op = optimizer.minimize(
loss=loss, global_step=tf.train.get_global_step()) loss=loss, global_step=tf.train.get_global_step())
# Provide an estimator spec for `ModeKeys.EVAL` and `ModeKeys.TRAIN` modes.
return tf.estimator.EstimatorSpec( return tf.estimator.EstimatorSpec(
mode=mode, mode=mode,
predictions=predictions_dict,
loss=loss, loss=loss,
train_op=train_op, train_op=train_op,
eval_metric_ops=eval_metric_ops) eval_metric_ops=eval_metric_ops)

19
tensorflow/examples/tutorials/estimators/abalone.py
View File

@ -87,25 +87,30 @@ def model_fn(features, labels, mode, params):
# Reshape output layer to 1-dim Tensor to return predictions # Reshape output layer to 1-dim Tensor to return predictions
predictions = tf.reshape(output_layer, [-1]) predictions = tf.reshape(output_layer, [-1])
predictions_dict = {"ages": predictions}
# Provide an estimator spec for `ModeKeys.PREDICT`.
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(
mode=mode,
predictions={"ages": predictions})
# Calculate loss using mean squared error # Calculate loss using mean squared error
loss = tf.losses.mean_squared_error(labels, predictions) loss = tf.losses.mean_squared_error(labels, predictions)
optimizer = tf.train.GradientDescentOptimizer(
learning_rate=params["learning_rate"])
train_op = optimizer.minimize(
loss=loss, global_step=tf.train.get_global_step())
# Calculate root mean squared error as additional eval metric # Calculate root mean squared error as additional eval metric
eval_metric_ops = { eval_metric_ops = {
"rmse": tf.metrics.root_mean_squared_error( "rmse": tf.metrics.root_mean_squared_error(
tf.cast(labels, tf.float64), predictions) tf.cast(labels, tf.float64), predictions)
} }
optimizer = tf.train.GradientDescentOptimizer( # Provide an estimator spec for `ModeKeys.EVAL` and `ModeKeys.TRAIN` modes.
learning_rate=params["learning_rate"])
train_op = optimizer.minimize(
loss=loss, global_step=tf.train.get_global_step())
return tf.estimator.EstimatorSpec( return tf.estimator.EstimatorSpec(
mode=mode, mode=mode,
predictions=predictions_dict,
loss=loss, loss=loss,
train_op=train_op, train_op=train_op,
eval_metric_ops=eval_metric_ops) eval_metric_ops=eval_metric_ops)